SOCIAL INFRASTRUCTURE AND LONG-RUN ECONOMIC
7.5 EMPIRICAL EVIDENCE
Our simple theoretical framework for analyzing investments has a number of general predictions. A country that attracts investments in the form of capital for businesses, technology transfer from abroad, and skills from individuals will be one in which the
• institutions and laws favor production over diversion,
• economy is open to international trade and competition in the global marketplace, and
• economic institutions are stable.
165 E M P I R I CAL E V I D E N C E
A good social infrastructure encourages domestic investment by fi rms in physical capital (factories and machines), investment by foreign entrepreneurs that may involve the transfer of better technologies, and the accumulation of skills by individuals. Furthermore, such an envi- ronment encourages domestic entrepreneurship; individuals look for better ways to create, produce, or transport their goods and services instead of looking for more effective ways to divert resources from other agents in the economy.
What empirical evidence supports these claims? Ideally, one would like empirical measures of the attributes of an economy that encourage the various forms of investment. Then, one could look at the economies of the world to see if these attributes are associated with high rates of investment and successful economic performance.
As the research literature on the role of these attributes has grown, so has the number of indices measuring them. The World Bank Governance Indicators Project, produced by Daniel Kaufmann, Aart Kraay, and Mas- simo Mastruzzi (2010), collects a range of indices from various sources related to accountability of politicians, political stability, government effectiveness, regulatory quality, the extent of the rule of law, and the control of corruption. They provide a summary measure for each of those six main areas, and we take the average of these six from the year 2008 to create an index of social infrastructure. This average index is closely correlated with different measures examined in Hall and Jones (1999);
Sachs and Warner (1995); Acemoglu, Johnson, and Robinson (2001);
Easterly and Levine (2003); and many others. Our index is normalized so that a value of one represents the best existing social infrastructure and a value of zero represents the worst.
Even though we have an index, simply showing that this is corre- lated with rates of investment and TFP is not suffi cient to prove that social infrastructure matters. The problem is one of casuality. It may be that a better social infrastructrure leads to higher investment rates, but it may also be the case that a high-quality social infrastructure is a luxury that countries with high investment rates (and hence high incomes) can afford. If investment rates drive social infrastructure, then we are back to square one, trying to explain differences in investment rates. Econo- mists have struggled with this empirical problem, and there is some relatively recent research that provides some assurance that it is social infrastructure that is in fact causing differences in economic outcomes.
Without getting into the technicalities of these estimates, the basic idea is similar to the “experiments” of Mancur Olson (1996) mentioned in the introduction. We need to fi nd situations in which the social infrastrure was changed exogenously in countries, without any direct infl uence from their investment rates or TFP levels. Acemoglu, Johnson, and Robinson (2001) take the colonization of countries around the world by Europe- ans as this kind of experiment. They fi nd that the exogenous differences in social infrastructure put in place by colonists had effects on income per capita even after the colonies obtained their independece much later.
Melissa Dell (2010) examines areas in Peru and Bolivia that were part of the mita, the forced labor system used by the Spanish from 1573 to 1812.
She compares these to areas outside of the mita but sharing a similar cul- tural and geographic background, and she fi nds that the mita led to much lower investment in public goods and poor economic outcomes. These studies provide us with our best evidence that social infrastructure does, in fact, drive differences in rates of investment. So while we will simply plot correlations in what follows, we can be fairly confi dent that much of what we observe represents the effect of social infrastructure.
Figures 7.1 and 7.2 plot investment as a share of GDP and aver- age educational attainment against this index of social infrastructure.
These fi gures show a relationship between social infrastructure and factor accumulation: countries with a good social infrastructure tend to have much higher investment rates in both physical and human capital. In countries where the social infrastructure allows investors to earn appropriate returns on their investments, fi rms and workers invest heavily in capital and skills.
This reasoning suggests a possible explanation of the stylized fact related to migration that we discussed in Chapter 1 (Fact 7). Recall that standard neoclassical theory suggests that rates of return are directly related to scarcity. If skilled labor is a scarce factor in developing econ- omies, the return to skill in these economies should be high, and this should encourage the migration of skilled labor out of rich countries and into poor countries. Empirically, however, the opposite pattern seems to occur. The explanation suggested here reverses this reasoning.
Suppose that, at least to a fi rst approximation, rates of return to skill are equalized by migration across countries. The stock of skills in devel- oping countries is so low because skilled individuals are not allowed to earn the full return on their skills. Much of their skill is wasted by
167 E M P I R I CAL E V I D E N C E
diversion—such as the payment of bribes and the risk that the fruits of their skill will be expropriated.5
Finally, Figure 7.3 plots the TFP level against social infrastructure.
Recall from Chapter 3 that some countries get much more output from their inputs (capital and skills) than do other countries. This is refl ected in differences in TFP across countries. Figure 7.3 shows that these differences are also related to social infrastructure. To see why this might be the case, consider a simple example in which individuals can choose to be either farmers or thieves. In the economy of Cornucopia, government policies strongly support production, no one is a thief, and society gets the maximum amount of output from its resources. On the
FI G U R E 7.1
UNDERSTANDING DIFFERENCES IN INVESTMENT RATESSOURCE: Author’s calculation using data from Appendix C and Kaufmann, Kraay, and Mastruzzi (2010)
ARG
AUS AUT
BDI
BEL BFABEN
BGD
BOL
BRA
BRB BWA
CAF
CAN CHE CHL
CHN
CIV COGCMR
COL COM
CPV
CRI
CYP
DNK DOM
DZA ECU
EGY
ESP
ETH
FJI FIN
FRA GAB
GBR GHA
GIN GMB
GNB GNQ
GRC GTM
HKG
HND
HTI
IDN IND
IRL
IRN ISR ISL
ITA JAM
JOR
JPN
KEN
KOR
LKA LSO
LUX MAR
MDG MEX MLI MOZ MRT
MUS
MWI MYS
NAM
NER
NGA
NIC
NLD NOR NPL
PAK PAN NZL
PER PHL
PNG PRI
PRT
PRY
ROM
RWA SEN
SGP
SLV
SWE TCD SYR
TGO
THA
TTO TUR
TWN TZA
UGA
URY
USA VEN
ZAF ZAR
ZMB ZWE
0.00 0.10 0.20 0.30 0.40 0.50
Investment share of GDP, 1988–2008
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Social infrastructure
5Migration restrictions could then explain the observed pattern that skilled labor migrates from developing countries to developed countries when it has the opportunity.
other hand, in the economy of Kleptocopia, whose policies do not sup- port production, thievery is an attractive alternative. Some individuals spend their time stealing from farmers. Thus some of the farmers’ time that might have been spent farming must be used to guard the crops against thieves. Similarly, some capital that might have been used for tractors is used for fences to keep out the thieves. The economy of Cor- nucopia gets much more output from its farmers and capital than does the economy of Kleptocopia. That is, Cornucopia has higher TFP.
This reasoning can help us rewrite the aggregate production func- tion of an economy, like that used in Chapter 6 in equation (6.3), as
Y = IKa(hL)1-a,
where I denotes the infl uence of an economy’s social infrastructure on the productivity of its inputs. With this modifi cation, we now have a complete theory of production that accounts for the empirical results documented in Chapter 3. Economies grow over time because new capital goods are
FI G U R E 7.2
UNDERSTANDING DIFFERENCES IN SKILL ACCUMULATIONSOURCE: Author’s calculation using data from Appendix C and Kaufmann, Kraay, and Mastruzzi (2010)
ARG
AUS
AUT
BDI
BEL
BEN BGD
BOL
BRA
BRB BWA
CAF
CAN
CHE CHL
CHN
CIV COGCMR
COL
CRI
CYP
DNK
DOM ECUDZA
EGY
ESP FIN
FJI
FRA
GAB
GBR
GHA
GMB
GRC
GTM
HKG
HND HTI
IDN IND
IRL
IRN
ISL ISR
JAM ITA JOR
JPN
KEN
KOR LKA
LSO
LUX
MAR MEX
MOZMLI MRT
MUS
MWI
MYS
NAM
NER NIC
NLD NOR
NPL
NZL
PAK
PAN PHL PER
PNG PRY PRT
ROM
RWA SEN
SGP SLV
SWE
SYRTGO
THA TTO
TUR
TWN
TZA UGA
URY
USA
VEN
ZAF ZMB
ZWE
0 2 4 6 8 10 12 14
Average years of schooling, 2008
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Social infrastructure
169 M I S A L L O C ATI O N A N D P R O D U CTI V IT Y
invented and the agents in the economy learn to use the new kinds of capital (captured by h). However, two economies with the same K, h, and L may still produce different amounts of output because the economic environments in which those inputs are used to produce output differ.
In one, capital may be used for fences, security systems, and pirate ships, and skills may be devoted to defrauding investors or collecting bribes. In another, all inputs may be devoted to productive activities.